Optimized convolutional neural network for the classification of lung cancer

Convolutional neural networks (CNN) have made significant strides in the field of image processing recently by solving a variety of previously unsolvable issues. However, the efficacy of this system depends on the selected hyper-parameters, and it is hard to physically adjust these hyper-parameters. As a result, an optimized convolution neural network is suggested in this study and is then employed to identify the kind of lung cancer. By employing an appropriate encoding strategy, the approach has used a gray wolf optimization algorithm to optimize hyper-parameters of CNN. By contrasting the model's performance with that of conventional CNN on the NIH/NCI Lung Image Database Consortium data set, the model's efficacy is confirmed. According to simulation findings, the suggested model can generate testing accuracy up to 98.21%, which is higher than CNN. Similarly, the suggested model's testing loss is around 0.10%, less than CNN. The test data conclusively show that the suggested model outperforms the conventional CNN.